3b1/7300 and NetHack3.4

[Kevin Darcy]

In article <2445 at flatline.UUCP> erict at flatline.UUCP (J. Eric Townsend) writes:
>
>
>Well, it seems 3.4 has finally outgrown the 3b1.  Even in super-stupid
>mode, my machine wouldn't compile w/out givng a 'too many defines'
>error.
>If you've gotten it to run on a stock 3b1, let me know how, puhleeze?
>For my curiosity's sake...

The 'too much defining' message (which is what I assume you mean, since that's
the message we got) means that the preprocessor is overflowing its symbol
space, which is not hard to do with Nethack.  I first saw the problem on a 3B2 
at work when we started installing NH 3.0, and we fixed it on that machine by 
getting our hands on a more robust version of the preprocessor.  Since I 
couldn't get this for my 3B1 at home, I gritted my teeth and #define'd 
"STUPID_CPP".  Now that the developers have gone and pushed it over even this
limit, mon.c and lev_comp.y now sing that old irritating song "too much 
defining", and more kludges are necessary.  A relatively painless (and tested,
by yours truly) workaround is to comment out or delete the inclusion of
"artifact.h" in mon.c (they go to the trouble of explicitly #include'ing that
file, and then don't use it), and to #define GOLD_H in lev_comp.y, which has
the effect of blocking that include file, and all of its unnecessary #define's.
Then sit back and make.  No muss, no fuss... at least until the next patch
level :-(   

-------------------------------------------------------------------------------
Opinions? What are those? I merely convey irrefutable truths...

 Kevin Darcy, Asst. Unix Systems Admin. | +1 313 948 4863
 kevin at cfctech.UUCP                     | Chrysler Financial Corp. (THE BEST)
 kevin%cfctech.uucp at mailgw.cc.umich.edu | MIS, Technical Services
 {sharkey|mailrus}!cfctech!kevin        | 27777 Franklin, Southfield, MI 48034
-------------------------------------------------------------------------------